Hands on with Gemini 2.5 Pro: why it might be the most useful reasoning model yet
1 min read
Summary
Google’s latest language model, Gemini 2.5 Pro, has been overshadowed by the Studio Ghibli AI image storm, delaying the release of OpenAIs’ new GPT-4.
However, the model is impressive and currently could be the best reasoning model with its long context window and output length.
It can process up to 1 million tokens, allowing it to fit multiple long documents and entire code repositories into the prompt.
Its long context window also allows for extended conversations and has performed well in multimodal prompts, extracting key information to create flowcharts and SVG graphics from a text description.
It can also perform useful data analysis, identifying stocks to invest in and calculating the value of a portfolio.
Google has presented Gemini 2.5 Pro as its most intelligent AI model and while it is only available in reasoning mode, its capabilities mean it could put Google at the forefront of the generative AI race.